30 



knowledge, it appears legitimate to assume their relationship to it. 

 In all the five specimens alluded to, there are more or less perfect evi- 

 dences of four distinct integuments, and of a large cavity, which is 

 in all filled with carbonate of lime and magnesia ; these minerals, I 

 presume, having replaced the albumen and embryo of the seed. 



The general form of the perfect fruit is an elongated ovoid (rather 

 larger than a hazel nut), of which the broader or lower end presents 

 the point of attachment, while the upper or smaller end is produced 

 into a straight, conical, truncated rostrum or beak, which is per- 

 forated by a straight longitudinal canal. The exterior integument 

 is very thick and cellular, and was no doubt once fleshy ; it alone is 

 produced beyond the seed and forms the beak ; its apex I assume to 

 have been that of the primine of the ovule, and its cavity the exo- 

 stome. The second coat appears to have been much thinner, but 

 hard and woody or bony; it is impervious at the apex; is also ovoid, 

 and sessile by its broad base within the outer integument, with 

 which it is perhaps adherent everywhere except at the apex. This 

 is marked by three angles or ridges, and being that alone which 

 (owing to its hard nature) commonly remains in the fossil state, has 

 suggested the name of Trigonocarpon. Within this are the third 

 and fourth coats, both of which are very delicate membranes ; one 

 appears to have been in close apposition with the inner wall of the 

 second integument, and the other to have surrounded the albumen. 

 These are now separated both from one another, and from the inner 

 wall of the cavity, by the shrinking of the contents of the latter, and 

 the subsequent infiltration of water charged with mineral matter. I 

 may remark, however, that these two membranes may be due to the 

 separation of one into two plates, in which case the original one was 

 formed of several layers of cells. Hitherto I have not been able to 

 trace any organized structure within the cavity of the fruit, and its 

 real nature therefore remains doubtful. It is only from the strong 

 resemblance, in structure, appearance and superposition, which these 

 integuments present to those of Taxoid coniferae, that I assume their 

 probable relationship. Salisburia, especially, has the same ovoid 

 fruit, sessile by its broader end, and its outer coat is perfectly ana- 

 logous, being thick, fleshy, and perforated at its apex by a longitu- 

 dinal canal (the exostome of the ovule) ; within this is a perfectly 

 similar, woody, two or thret angled, impervious integument, form- 



